What fits, in the order you meant
Takes your backlog as prioritised, protects a reserve for the support work that always arrives, and tells you what was left out and why — without rearranging your priorities to pack better.
Filling a sprint against a real capacity
Takes your backlog in the order you prioritised it, respects dependencies, and tells you what did not fit and why.
Backlog, in priority order
In the sprint
Did not make it
How selection works
Items are considered in the order you listed them. An item is taken if it fits in the remaining capacity and all of its dependencies have already been taken.
Nothing is reordered. A knapsack optimiser would fit marginally more work in by rearranging your priorities, and that is not a trade anyone actually wants — the highest-priority item does not become less important because a cheaper one would pack better.
The reserve is subtracted from capacity before selection, so it is protected rather than merely intended.
Everything excluded is listed with the reason, because "these did not fit" is only useful if you can see whether it was size, sequence or a dependency.
The reserve is the point
Every team knows support and interruption will arrive, and most sprint plans are built as though they will not. The result is a sprint that is full on day one and late by day three, with the unplanned work done anyway and the committed work quietly dropped.
Reserving capacity up front changes what gets committed rather than what gets done. The work arrives either way; the only question is whether the plan admitted it would.
A full sprint is a fragile sprint
Filling to 95% or more of capacity leaves nothing for the first surprise, and there is always a first surprise. The teams that hit their sprint goals consistently are usually the ones that commit to noticeably less than they could.
This is unintuitive to anyone who reads a sprint plan as a productivity target. It stops being unintuitive the moment you have watched a fully-loaded sprint fail on something nobody could have predicted.
What this deliberately will not decide
Whether to split an item that nearly fits. Splitting is a conversation about what a smaller version would deliver, and there is no arithmetic that makes it for you — an item split badly produces two things neither of which is useful alone.
The tool flags the near-miss and stops there, which is the right place for a calculator to stop.